Skip to main content

Editing an AI-Generated Scenario

Conversational phishing scenario

Updated over a month ago

To customize the phishing email and include personalized elements, click Duplicate on a scenario card.

After completing the scenario details step, click Continue.

Phishing scenario creation screen showing scenario name, language, difficulty, and the highlighted Continue button.


1 – Objectives

  • Adapt the AI-generated email to your organization

  • Add dynamic elements using tokens

  • Identify and highlight Red Flags in the training page

  • Perform advanced customization by editing the generated structure


2 – Prerequisites

  • Admin permissions to modify scenarios

  • A duplicated AI scenario with the details step already completed


3 – Edit the general email settings

3.1 – Access the AI Builder

Click AI Builder to open the AI-based email editor

Email customization screen with the AI Builder option selected, enabling automatic generation of phishing emails.

3.2 – Edit the main fields

Customize:

  • Sender’s name

  • Sender’s email address

3.3 – Configure the sender domain

  • Select a sending domain from the dropdown

To add a custom domain:

  • Go to Settings

  • Open Phishing Domains

  • Click + Add New Phishing Domain

List of available phishing domains with the Add New Phishing Domain button highlighted.


4 – Use Red Flags

4.1 – Enable or disable a Red Flag

  • Check the Red Flag box to mark a field as suspicious

  • Uncheck it if no alert is needed

Red Flag icon enabled on the sender address field to highlight a suspicious element.

4.2 – Display Red Flags in the training page

Checked fields will appear highlighted in red in the Just-in-Time Training page

To edit the explanatory messages:

Go to Settings → Just-in-Time Training Page

⚠️ Warning
Red Flags only appear in training campaigns.
For the difference between training and assessment campaigns, click here.


5 – Edit the email body

5.1 – Adjust the AI generation parameters

You can adjust several parameters used by the AI to generate unique phishing emails:

  • Email subject

  • Job title (who the email is “from”)

  • Tone of the email

  • Main attack prompt describing the scenario logic

  • Optional custom signature

5.2 – Edit the email subject

  • Edit the Email Subject on the right side of the screen

  • Use the button to generate a subject automatically via AI

Email Subject field featuring the AI-generation “sparkle” button.

5.3 – Select the sender’s job role

  • Choose a Job Title to define the identity used in the generated email

  • Provide additional context to strengthen the realism of the attack

Job title dropdown opened, showing IT, HR, Accounting, Marketing, etc.

5.4 – Adjust the tone of the email

Move the Tone slider to choose between:

  • Friendly

  • Cordial

  • Professional

Tone slider positioned between Friendly and Professional, with “Cordial” selected.

5.5 – Write or adjust the AI prompt

  • Use the predefined prompts available in the dropdown, or

  • Write your own custom prompt

Pretext prompt field showing a custom attack script related to security patch installation.

📚 Recommendations for writing prompts

  • Use short sentences

  • Focus on action verbs

  • Build on the context defined by the subject, the Job Title and the Tone.

5.6 – Add a signature

  • Click Custom Agent Signature to add a signature block

You may:

  • Add a real employee signature for realism

  • Insert an invented signature to maintain an attacker-style identity

Custom signature editor opened, allowing definition of a personalized signature.


6 – Choose the type of AI scenario

Two modes are available:

  • Conversational scenarios

  • Traditional (non-conversational) scenarios

6.1 – Conversational scenarios

6.1.1 – Objectives & benefits

Conversational phishing scenarios allow you to:

  • Simulate attack techniques representing ~80% of today’s phishing attempts

  • Show how attackers build trust and bypass technical defenses

  • Expose employees to AI-enabled attack patterns

  • Provide highly realistic training through multi-email conversations

6.1.2 – Practical setup

AI-generated conversational scenarios simulate an ongoing exchange between the attacker and the employee.

  • The conversation can span multiple emails

  • Each reply from the user influences the next AI-generated message

AI scenario test window showing a simulated message exchange between the attacker agent and the target.

To test the scenario:

  • Click Test your AI scenario

"Test AI Scenario” option highlighted to launch the simulation.

6.2 – Traditional (non-conversational) scenarios

Selecting a non-conversational AI scenario means:

  • Each employee receives a unique email

  • No multi-email conversation occurs

  • User replies are not considered

  • All emails still include a phishing link

To use this mode:

  • Disable the option Enable fully automated conversational phishing mode

Did this answer your question?